experience, most turning points in our level are permanent bench marks no more than 40 m apart throughout the line, and permanently marked instrument points are exactly midway between bench marks. The relatively short, balanced sights minimize systematic errors from refraction and collimation, respectively, and assure that the rod images are sufficiently large to be read accurately. Our leveling is always double-run using a shaded precision automatic level and strut-supported Invar leveling rods. The same rod is always placed on the same bench mark in each successive survey, and our rods are annually calibrated every 10 cm along their entire length.
Perhaps because only a few investigators have been leveling in the near field for a short time in limited parts of the world, few reports of significant results are available in contrast to those from broad-scale leveling. A review of vertical preseismic, coseismic, and postseismic slip associated with U.S. earthquakes is given by Reilinger and Brown (1981). Vertical afterslip was indicated but not authenticated in two New Zealand earthquakes—the Murchison earthquake of 1929 (Henderson, 1937) and the Napier earthquake of 1931 (Henderson, 1933). Otherwise vertical afterslip has been documented in only four earthquakes. Sylvester and Pollard (1975) found that afterslip amounted to less than 1 percent of the coseismic vertical separation in the year following the 1971 San Fernando (M=6.4) earthquake; Sharp and Lienkaemper (1982) found that 14-cm afterslip, nearly equal to the coseismic slip of 16 cm, occurred in the 10 weeks following the 1979 Imperial Valley (M=6.5) earthquake; afterslip following the Alaskan (M=8.3) earthquake of 1964 reached 0.55 m over 10 years (Brown et al., 1977; Prescott and Lisowski, 1977); Lensen and Suggate (1968) measured 12 mm of vertical afterslip in 2 months across the Inangahua Fault (New Zealand) after about 1 m of vertical separation in the 1968 Inangahua (M=7) earthquake. Near-field leveling revealed 54 µrad of crustal tilt 6 months before the Imperial Valley earthquake (Sharp and Lienkaemper, 1982), and significant tilt has also occurred before some Chinese earthquakes (Mei, 1984; Zhu et al., 1984). Previous reports of regional tilt before Japanese earthquakes have been attributed to refraction errors in the leveling (Mogi, 1984).
Vertical movement across normal faults of the Asal-Ghoubbet rift, Dijoubti, East Africa, is regarded as aseismic creep by Ruegg et al. (1984), but vertical creep has not been observed elsewhere in spite of specific searches for it, especially in California (Sylvester, 1982).
Tilt of the Earth’s surface has been observed prior to earthquakes by broad-scale leveling (e.g., Bendefy,